ACCURACY OF PROCALCITONIN FOR OUTCOME PREDICTION IN UNSELECTED POSTOPERATIVE CRITICALLY ILL PATIENTS

The importance of postoperative procalcitonin (PCT) measurements for outcome prediction is currently controversial. Conflicting results have been obtained for patients after polytrauma, sepsis, peritonitis, or cardiac surgery and may result from incomplete adjustment for important confounders or from nonlinear PCT effects. We retrospectively analyzed the association of PCT concentration with postoperative mortality, morbidity, and length of stay in an unselected series of 220 consecutive patients who required postoperative intensive care unit therapy or surveillance. Biochemical markers were measured on the first day after intensive care unit admission. Results were adjusted for various confounding variables (Acute Physiology and Chronic Health Evaluation II score, underlying disease), and test accuracy was evaluated by receiver operating characteristic statistics. We found a significant nonlinear, logarithmic association between PCT concentration and outcome. After adjustment for relevant covariates, PCT was an independent determinant of mortality, combined mortality/morbidity,and postoperative hospital length of stay in survivors. At mortality analysis, the predictive power of PCT was superior to that of Acute Physiology and Chronic Health Evaluation II score and of IL-6 (optimal cutoff point, 1.44 ng/mL; sensitivity, 80.8%; specificity, 80.4%). The use of PCT was comparable to that of other prognostic markers when combined mortality/morbidity were examined. Our results suggest that PCT may deserve further testing as a prognostic tool in unselected, critically ill, surgical patients.